PSOL

We have updated the validation code for PSOL on ImageNet; CUB code and Training Code will be available soon;

This is the offical website for PSOL.

This package is developed by Mr. Chen-Lin Zhang (http://lamda.nju.edu.cn/zhangcl/) and Mr. Yun-Hao Cao (http://lamda.nju.edu.cn/caoyh/). If you have any problem about the code, please feel free to contact Mr. Chen-Lin Zhang (zhangcl@lamda.nju.edu.cn). The package is free for academic usage. You can run it at your own risk. For other purposes, please contact Prof. Jianxin Wu (wujx2001@gmail.com).

If you find our package is useful to your research, please cite our paper:

Reference:

[1] C.-L. Zhang, Y.-H. Cao and J. Wu. Rethinking the Route Towards Weakly Supervised Object Localization . In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2020), Seattle, WA, USA, in press.

Requirements

The code needs PyTorch and OpenCV.

A GPU is optional for faster speed.

Other requirements are in requirements.txt. You can simply install them by pip install -r requirements.txt. We recommend to use Anaconda with virtual environments.

Usage

prepare datasets:

Training:

Soon.

Validation:

First you should download the ImageNet validation set, and corresponding annotation xmls. We need both validation set and annotation xmls in PyTorch format. Please refer to ImageNet example for details.

Training

Soon.

Testing

We have provided some pretrained models. If you want to directly test it, please download it.

First you need to modify some variables in PSOL_inference.py:

Line 229-231: Environment variables.

Line 264-268: Root folder for validation image files and annotation files.

Then you can run:

python PSOL_inference.py --loc-model {$LOC_MODEL} --cls-model {$CLS_MODEL} {--ten-crop}

$LOC_MODEL represents the localization models we support now.

Options:

densenet161: DenseNet161

vgg16: VGG16

vgggap: replace all fc layers in VGG16 with a gap layer and a single fc layer

resnet50: ResNet50

inceptionv3: coming soon (because of the torchvision issue)

$CLS_MODEL represents the classification models we support now. Options:

densenet161: DenseNet161

vgg16: VGG16

resnet50: ResNet50

inceptionv3: InceptionV3

dpn131: DPN131

efficientnetb7: EfficientNetB7

Please note that for SOTA classification models, you should change the resolutions in cls_transforms (Line248-Line249).

--ten-crop means that the classification models are averaged from ten crops.

Then you can get the final result.